Mechanism Of Electro-Fenton Chemical Mechanical Polishing For Single Crystal SiC

Owing to its excellent physical and chemical properties,single crystal SiC is widely used for various applications,such as 5G communications,military defence,new energy vehicles,LED lighting,consumer electronics and other fields.As an essential epitaxial substrate material,SiC requires an atomic-level ultra-smooth surface without processing subsurface damage and free residual stresses on the processed surface.However,due to the high hardness,high strength,and stable chemical properties of SiC materials,the application of single crystal SiC is limited by the difficulty and high cost of high-efficiency ultra-precision planarization processing.Currently,chemical mechanical polishing(CMP)is commonly used for the global planarisation polishing of semiconductor substrates;the chemical reaction process is crucial for CMP.Hydrating hydroxyl radicals(·OH)with strong oxidisability can effectively oxidise and corrode the SiC material.In this thesis,we propose the Electro-Fenton Chemical Mechanical Polishing(EF-CMP)method,which uses the electric field parameters of the Electro-Fenton reaction to control the generation rate and total·OH concentration during the CMP,to achieve a continuous,stable and controllable generation of·OH;and accelerate the SiC substrate oxidation and corrosion to achieve the synergistic effect with the abrasive removal process,which in turn to improve the material removal rate(MRR)and flattening effect of SiC substrate during the CMP,and to explore a high-efficiency and high-precision ultra-smooth surfaces planarization and polishing technology.Firstly,the Electro-Fenton reaction conditions and the reaction mechanism were optimized and studied by orthogonal and single-factor experiments.The results show that the initial H₂O₂ concentration,the Fe₃O₄ concentration and the voltage have the most significant effects on the·OH generation in the polishing slurry;The maximum and total·OH concentration generated by the Electro-Fenton reaction first increased and then decreased with increasing initial H₂O₂ concentration,Fe₃O₄ concentration,and voltage.In the Electro-Fenton reaction,controlling the electric field parameters,on the one hand,can increase Fe²⁺concentration(the growth rate is about 97.49%)in the reaction solution and enhance the catalytic effect of the reaction.On the other hand,it also generates a small amount of H₂O₂ in situ on the cathode surface(the generated concentration is about 42.84%of the initial H₂O₂ concentration),compensating for H₂O₂ consumption.Furthermore,the Pt anode oxidises water to generate a small amount of·OH(the generated concentration is about 13.75%of the total·OH concentration),and these combined effects promote·OH generation.Additionally,the total·OH concentration achieved in the Electro-Fenton reaction was 48.69%greater than that produced by the Fenton reaction.Secondly,the relationship between the oxidation activity(total·OH concentration)of the polishing slurry,coefficient of friction(COF),the thickness of the oxide layer,and polishing effect was systematically studied through immersion oxidation experiments,surface friction and wear experiments,and CMP experiments.And the mechanism of the EF-CMP of SiC was also analysed.The results indicated that applying an external field voltage in the Fenton reaction could significantly enhance the oxidation activity of the polishing slurry.A higher voltage and H₂O₂concentration produce higher maximum and total·OH concentration,stronger oxidation activity,stronger chemical reaction on the SiC surface,and better polishing effect.When the H₂O₂ concentration was 5.0 wt%,and the applied voltage increased from 0.0 V to 1.5 V and 3.0 V,the oxidation activity increased by 133.47%and196.24%;COF increased by 9.05%and 13.36%;and MRR increased by 35.00%and54.23%,respectively.At a high voltage(3.0 V)and high H₂O₂ concentration(7.5wt%),the polishing slurry exhibited the most potent oxidation activity,and the oxidised SiC surface was the roughest;the COF and MRR were the highest,and the polishing effects was the best.Compared to the case with no electric field and low H₂O₂ concentration(5.0 wt%),the oxidation activity,COF,and MRR increased by256.92%,18.53%,and 72.31%,respectively.Thirdly,based on the modified Preston equation,the effects of the EF-CMP process parameters on the MRR and Ra after polishing of SiC and their relationship were studied using factor analysis.Combined with the experimental data,the adaptive momentum(Adam)optimisation algorithm is used to establish a prediction model based on the BP neural network between the process parameters,MRR,and Ra after polishing for SiC during EF-CMP.The results indicated that by increasing the external electric field voltage and H₂O₂ concentration,the MRR gradually increased and the Ra after polishing decreased.Increasing the Fe₃O₄ concentration,the MRR initially increased and then decreased,and the Ra initially decreased and then increased.Increasing the abrasive size,abrasive concentration,polishing pressure,and polishing speed,the MRR continuously increased,but the Ra initially decreased and then increased.The established prediction model based on the experimental results could accurately predict the relationship between the process parameters,MRR and Ra after polishing in EF-CMP(relative prediction error of less than 10%),which solving the difficulty of traditional theoretical models to accurately describe the complex non-linear mapping relationship between process parameters and experimental targets,which solving the prediction problem of non-linear mapping relationship between multiple process parameters on a single-evaluation index(MRR or surface roughness Ra).Finally,an orthogonal experiment was used to study the effects of external electric field voltage(A),initial H₂O₂concentration(B),abrasive concentration(C)and polishing pressure(D)and their interactions on MRR,Ra,maximum peak-valley roughness(Rt)and TTV to optimise the EF-CMP process parameters of single crystal SiC;and to construct a regression equation model.The entropy value method was also used to assign weights for each evaluation index,and grey relational analysis was used to transform the optimisation problem of multi-evaluation indexes into a single-index optimisation problem regarding the grey relational grade(GRG),and finally realize the linear mapping relationship problem between multiple process parameters and multi-evaluation indexes using the established regression equation model to obtain the polishing process parameters with a strong,comprehensive performance.The results show that the contribution of the factors[D,B,A,(A×D),(A×B)],[A,B,C,D],[A,B,D]and[C,A]to the MRR,Ra,Rt and TTV were as high as 99.84%,90.74%,63.93%and 74.71%.The process parameter optimised by the grey relational analysis resulted in a smooth surface with an MRR of 2357.740 nm/h,a Ra of 0.372 nm,an Rt of 6.346 nm,and a TTV of 3?m,which is 9.27%,9.17%,20.90%and 31.14%better than the comprehensive polishing performance obtained from the analysis under the single-evaluation indexes conditions,respectively.